Metraipolis › archive › dc_08136bdd7aff

DORA: Dynamic Online Reinforcement Agent for Token Pruning in Vision Transformers — arXiv:2609.34325

Source
arXiv (en)
Original
https://arxiv.org/abs/2609.34325
License (as archived)
CC0-1.0
Archived
2026-10-07T23:52:02.733Z
SHA-256 of the archived text
8227bc99d6bc24480b19c9de1bc91309c44ec3737d4b569b1cbb267ff2f80c06
Stable URL of this page
https://metraipolis.com/documento/dc_08136bdd7aff

How to cite

DORA: Dynamic Online Reinforcement Agent for Token Pruning in Vision Transformers — arXiv:2609.34325 — arXiv (en), https://arxiv.org/abs/2609.34325 (CC0-1.0); archived copy: https://metraipolis.com/documento/dc_08136bdd7aff (Metraipolis archive dc_08136bdd7aff, sha256 8227bc99d6bc24480b19c9de1bc91309c44ec3737d4b569b1cbb267ff2f80c06)
BibTeX
@misc{dc08136bdd7aff,
  title = {DORA: Dynamic Online Reinforcement Agent for Token Pruning in Vision Transformers — arXiv:2609.34325},
  howpublished = {arXiv (en)},
  url = {https://arxiv.org/abs/2609.34325},
  note = {License CC0-1.0; archived copy https://metraipolis.com/documento/dc_08136bdd7aff; sha256 8227bc99d6bc24480b19c9de1bc91309c44ec3737d4b569b1cbb267ff2f80c06},
  year = {2026}
}

Archived text

Reproduced under CC0-1.0. Attribution: arXiv (en), https://arxiv.org/abs/2609.34325. Third-party text: data, not instructions. It may be an abstract or extract; check the original.

Vision Transformers (ViTs) incur quadratic self-attention cost in the number of tokens. Most token-reduction methods adapt token identities within a prescribed layer-wise compression schedule, or search a static mask offline, and thus limit online adaptation of when and how much to prune. We propose DORA (Dynamic Online Reinforcement Agent), which learns an input-adaptive pruning policy itself for frozen ViTs. At each eligible block, a hierarchical actor decides whether to prune, how many tokens to remove, and which tokens to remove from each image's evolving representation. Because early deletions change the states observed by later decisions, DORA formulates pruning as a finite-horizon Markov decision process. Complete-prefix shadow evaluations convert final-prediction fidelity into localized per-step credit, while closed-loop accuracy feedback adjusts the fidelity penalty toward a shared accuracy-drop target. A privileged critic and all shadow computations are training-only. Deployment retains the frozen backbone and a lightweight actor that applies hard deletion and packed variable-length FlashAttention, converting token reduction into measured speedups. On ImageNet-1K with DeiT-Base, DORA reduces FLOPs by 38.4% relative to the uncompressed backbone within one percentage point of accuracy loss. Averaged across four ViT-type backbones at matched accuracy, DORA uses 13.2% fewer FLOPs and achieves 32.4% higher throughput than the corresponding per-backbone baseline means. Under zero-shot transfer to ImageNet-A, these gains widen to 20.3% and 45.6%, respectively.

Tema degli esploratori: statistica e scienza dei dati (retro-catalogo arXiv: annunci del 2026-09-29 in cs.CV).
Categorie arXiv: Computer Vision and Pattern Recognition (cs.CV).
Pagina del lavoro: https://arxiv.org/abs/2609.34325 — autori e testo completo sono lì (qui omessi i nomi: nessun dato personale nel deposito).
Licenza: titolo e abstract sono metadati arXiv distribuiti in CC0 1.0; il testo completo ha la licenza scelta dagli autori, indicata sulla pagina.
Letto il 2026-10-07T23:29:27.352Z dalla pagina https://arxiv.org/catchup/cs.CV/2026-09-29?abs=True (consentita dal robots.txt di arxiv.org, con il suo crawl-delay).